Hackathon Demo — Apify Agent Pipeline

Know what your
GPUs actually cost

Real-time energy monitoring, cost attribution, and carbon tracking for every GPU workload. Plus AI-powered prospect discovery to find your next customers.

Try Prospect PipelineView Dashboard

What we built

⚡

GPU Cost Attribution

Track energy spend per job, model, and team in real-time

🔍

Waste Detection

Identify idle GPUs and underutilized workloads automatically

🌱

Carbon Tracking

Monitor CO₂ emissions with grid-aware carbon intensity data

🤖

Prospect Pipeline

Apify-powered agents discover and enrich leads automatically

💰

Budget Alerts

Set spend limits and get notified before you blow your budget

📊

Green AI Index

Benchmark your efficiency against anonymous industry peers

Powered by Apify

Two-Agent Prospect Pipeline

We built two autonomous agents that use Apify actors to find, enrich, verify, and load potential customers into our CRM — all from a single button click.

1

Discovery Agent

Finds AI/ML companies on LinkedIn and extracts decision-maker contact info.

a
LinkedIn Company Search
Actor: harvestapi/linkedin-company-search
Searches by keywords like "GPU cloud", "AI infrastructure", "deep learning". Filters by company size and location. Returns company profiles with website, industry, and employee count.
b
Decision Maker Email Finder
Actor: snipercoder/decision-maker-email-finder
Takes each company's domain, finds CEO/founder/owner contacts with name, email, title, phone, and LinkedIn URL. Up to 3 leads per company.
2

Verification + Load Agent

Validates emails and loads verified prospects into our Supabase CRM.

a
Email Verification
Actor: snipercoder/email-validator
Checks DNS, MX records, and SMTP for each email. Marks valid vs. invalid so we only reach out to real addresses. ~$1/1K verifications.
b
CRM Load
Storage: Supabase PostgreSQL
Upserts prospects into our database with dedup on domain + email. Company info, contact details, verification status — ready for outreach.

Pipeline Flow

Keywords
GPU cloud, AI infra...
→
LinkedIn Search
harvestapi actor
→
Email Enrichment
snipercoder actor
→
Verification
email-validator
→
CRM Load
Supabase upsert
Agent 1: Discovery
Agent 2: Verify + Load
Try the Pipeline Live

Architecture

Agent (Python)

NVML probe, WAL, batched upload. Runs as systemd/K8s DaemonSet.

API (Next.js)

Ingest, dashboard, cost, carbon, benchmarks, prospect agents.

Apify Pipeline

LinkedIn search, email enrichment, verification, CRM load.